X-Git-Url: https://pintos-os.org/cgi-bin/gitweb.cgi?a=blobdiff_plain;f=src%2Flanguage%2Flexer%2Fvariable-parser.c;h=3b6b96ad0e2e2135d49de0ea82dcda9577be0300;hb=0fde6afee3c995bf264c24c438f43eeb58b859b5;hp=a45d384293ba9b6251433a60abd4d89480dbc9fa;hpb=55c55aa33d0f90d1b3b58f8b33b3fc54062c553e;p=pspp diff --git a/src/language/lexer/variable-parser.c b/src/language/lexer/variable-parser.c index a45d384293..3b6b96ad0e 100644 --- a/src/language/lexer/variable-parser.c +++ b/src/language/lexer/variable-parser.c @@ -1,5 +1,5 @@ /* PSPP - a program for statistical analysis. - Copyright (C) 1997-9, 2000, 2009, 2010, 2011, 2012 Free Software Foundation, Inc. + Copyright (C) 1997-9, 2000, 2009, 2010, 2011, 2012, 2020 Free Software Foundation, Inc. This program is free software: you can redistribute it and/or modify it under the terms of the GNU General Public License as published by @@ -46,26 +46,42 @@ #include "gettext.h" #define _(msgid) gettext (msgid) -static struct variable * var_set_get_var (const struct var_set *, size_t ); - +static struct variable *var_set_get_var (const struct var_set *, size_t); static struct variable *var_set_lookup_var (const struct var_set *, const char *); - static bool var_set_lookup_var_idx (const struct var_set *, const char *, size_t *); +static bool var_set_get_names_must_be_ids (const struct var_set *); + +static bool +is_name_token (const struct lexer *lexer, bool names_must_be_ids) +{ + return (lex_token (lexer) == T_ID + || (!names_must_be_ids && lex_token (lexer) == T_STRING)); +} +static bool +is_vs_name_token (const struct lexer *lexer, const struct var_set *vs) +{ + return is_name_token (lexer, var_set_get_names_must_be_ids (vs)); +} +static bool +is_dict_name_token (const struct lexer *lexer, const struct dictionary *d) +{ + return is_name_token (lexer, dict_get_names_must_be_ids (d)); +} /* Parses a name as a variable within VS. Sets *IDX to the variable's index and returns true if successful. On failure emits an error message and returns false. */ static bool parse_vs_variable_idx (struct lexer *lexer, const struct var_set *vs, - size_t *idx) + size_t *idx) { assert (idx != NULL); - if (lex_token (lexer) != T_ID) + if (!is_vs_name_token (lexer, vs)) { lex_error (lexer, _("expecting variable name")); return false; @@ -107,7 +123,8 @@ parse_variable (struct lexer *lexer, const struct dictionary *d) /* Parses a set of variables from dictionary D given options OPTS. Resulting list of variables stored in *VAR and the number of variables into *CNT. Returns true only if - successful. */ + successful. The dictionary D must contain at least one + variable. */ bool parse_variables (struct lexer *lexer, const struct dictionary *d, struct variable ***var, @@ -121,6 +138,12 @@ parse_variables (struct lexer *lexer, const struct dictionary *d, assert (cnt != NULL); vs = var_set_create_from_dict (d); + if (var_set_get_cnt (vs) == 0) + { + *cnt = 0; + var_set_destroy (vs); + return false; + } success = parse_var_set_vars (lexer, vs, var, cnt, opts); var_set_destroy (vs); return success; @@ -201,9 +224,9 @@ add_variable (struct variable ***v, size_t *nv, size_t *mv, "All variables in this variable list must have the " "same width. %s will be omitted from the list."), var_get_name ((*v)[0]), add_name, add_name); - else if ((pv_opts & PV_NO_DUPLICATE) && included[idx]) + else if ((pv_opts & PV_NO_DUPLICATE) && included && included[idx]) msg (SE, _("Variable %s appears twice in variable list."), add_name); - else if ((pv_opts & PV_DUPLICATE) || !included[idx]) + else if ((pv_opts & PV_DUPLICATE) || !included || !included[idx]) { if (*nv >= *mv) { @@ -344,7 +367,8 @@ parse_var_set_vars (struct lexer *lexer, const struct var_set *vs, lex_match (lexer, T_COMMA); } while (lex_token (lexer) == T_ALL - || (lex_token (lexer) == T_ID && var_set_lookup_var (vs, lex_tokcstr (lexer)) != NULL)); + || (is_vs_name_token (lexer, vs) + && var_set_lookup_var (vs, lex_tokcstr (lexer)) != NULL)); if (*nv == 0) goto fail; @@ -360,6 +384,22 @@ fail: return 0; } +char * +parse_DATA_LIST_var (struct lexer *lexer, const struct dictionary *d) +{ + if (!is_dict_name_token (lexer, d)) + { + lex_error (lexer, "expecting variable name"); + return NULL; + } + if (!dict_id_is_valid (d, lex_tokcstr (lexer), true)) + return NULL; + + char *name = xstrdup (lex_tokcstr (lexer)); + lex_get (lexer); + return name; +} + /* Attempts to break UTF-8 encoded NAME into a root (whose contents are arbitrary except that it does not end in a digit) followed by an integer numeric suffix. On success, stores the value of the suffix into *NUMBERP, @@ -429,6 +469,7 @@ parse_DATA_LIST_vars (struct lexer *lexer, const struct dictionary *dict, char *name1 = NULL; char *name2 = NULL; + bool ok = false; assert ((pv_opts & ~(PV_APPEND | PV_SINGLE @@ -456,36 +497,24 @@ parse_DATA_LIST_vars (struct lexer *lexer, const struct dictionary *dict, do { - if (lex_token (lexer) != T_ID - || !dict_id_is_valid (dict, lex_tokcstr (lexer), true)) - { - lex_error (lexer, "expecting variable name"); - goto exit; - } - if (dict_class_from_id (lex_tokcstr (lexer)) == DC_SCRATCH - && (pv_opts & PV_NO_SCRATCH)) + name1 = parse_DATA_LIST_var (lexer, dict); + if (!name1) + goto exit; + if (dict_class_from_id (name1) == DC_SCRATCH && pv_opts & PV_NO_SCRATCH) { msg (SE, _("Scratch variables not allowed here.")); goto exit; } - name1 = xstrdup (lex_tokcstr (lexer)); - lex_get (lexer); - if (lex_token (lexer) == T_TO) + if (lex_match (lexer, T_TO)) { - unsigned long int num1, num2; + unsigned long int num1, num2; int n_digits1, n_digits2; int root_len1, root_len2; unsigned long int number; - lex_get (lexer); - if (lex_token (lexer) != T_ID - || !dict_id_is_valid (dict, lex_tokcstr (lexer), true)) - { - lex_error (lexer, "expecting variable name"); - goto exit; - } - name2 = xstrdup (lex_tokcstr (lexer)); - lex_get (lexer); + name2 = parse_DATA_LIST_var (lexer, dict); + if (!name2) + goto exit; root_len1 = extract_numeric_suffix (name1, &num1, &n_digits1); if (root_len1 == 0) @@ -608,21 +637,20 @@ parse_mixed_vars (struct lexer *lexer, const struct dictionary *dict, assert (names != NULL); assert (nnames != NULL); - assert ((pv_opts & ~PV_APPEND) == 0); if (!(pv_opts & PV_APPEND)) { *names = NULL; *nnames = 0; } - while (lex_token (lexer) == T_ID || lex_token (lexer) == T_ALL) + while (is_dict_name_token (lexer, dict) || lex_token (lexer) == T_ALL) { if (lex_token (lexer) == T_ALL || dict_lookup_var (dict, lex_tokcstr (lexer)) != NULL) { struct variable **v; size_t nv; - if (!parse_variables (lexer, dict, &v, &nv, PV_NONE)) + if (!parse_variables (lexer, dict, &v, &nv, pv_opts)) goto fail; *names = xnrealloc (*names, *nnames + nv, sizeof **names); for (i = 0; i < nv; i++) @@ -630,12 +658,12 @@ parse_mixed_vars (struct lexer *lexer, const struct dictionary *dict, free (v); *nnames += nv; } - else if (!parse_DATA_LIST_vars (lexer, dict, names, nnames, PV_APPEND)) + else if (!parse_DATA_LIST_vars (lexer, dict, names, nnames, PV_APPEND | pv_opts)) goto fail; } if (*nnames == 0) goto fail; - + return true; fail: @@ -669,9 +697,139 @@ parse_mixed_vars_pool (struct lexer *lexer, const struct dictionary *dict, struc return retval; } +/* Frees the N var_syntax structures in VS, as well as VS itself. */ +void +var_syntax_destroy (struct var_syntax *vs, size_t n) +{ + for (size_t i = 0; i < n; i++) + { + free (vs[i].first); + free (vs[i].last); + } + free (vs); +} + +/* Parses syntax for variables and variable ranges from LEXER. If successful, + initializes *VS to the beginning of an array of var_syntax structs and *N_VS + to the number of elements in the array and returns true. On error, sets *VS + to NULL and *N_VS to 0 and returns false. */ +bool +var_syntax_parse (struct lexer *lexer, struct var_syntax **vs, size_t *n_vs) +{ + *vs = NULL; + *n_vs = 0; + + if (lex_token (lexer) != T_ID) + { + lex_error (lexer, _("expecting variable name")); + goto error; + } + + size_t allocated_vs = 0; + do + { + if (allocated_vs >= *n_vs) + *vs = x2nrealloc (*vs, &allocated_vs, sizeof **vs); + struct var_syntax *new = &(*vs)[(*n_vs)++]; + *new = (struct var_syntax) { .first = ss_xstrdup (lex_tokss (lexer)) }; + lex_get (lexer); + + if (lex_match (lexer, T_TO)) + { + if (lex_token (lexer) != T_ID) + { + lex_error (lexer, _("expecting variable name")); + goto error; + } + + new->last = ss_xstrdup (lex_tokss (lexer)); + lex_get (lexer); + } + } + while (lex_token (lexer) == T_ID); + return true; + +error: + var_syntax_destroy (*vs, *n_vs); + *vs = NULL; + *n_vs = 0; + return false; +} + +/* Looks up the N_VS var syntax structs in VS in DICT, translating them to an + array of variables. If successful, initializes *VARS to the beginning of an + array of pointers to variables and *N_VARS to the length of the array and + returns true. On error, sets *VARS to NULL and *N_VARS to 0. + + For the moment, only honors PV_NUMERIC in OPTS. */ +bool +var_syntax_evaluate (const struct var_syntax *vs, size_t n_vs, + const struct dictionary *dict, + struct variable ***vars, size_t *n_vars, int opts) +{ + assert (!(opts & ~PV_NUMERIC)); + + *vars = NULL; + *n_vars = 0; + + size_t allocated_vars = 0; + for (size_t i = 0; i < n_vs; i++) + { + struct variable *first = dict_lookup_var (dict, vs[i].first); + if (!first) + { + msg (SE, _("%s is not a variable name."), vs[i].first); + goto error; + } + + struct variable *last = (vs[i].last + ? dict_lookup_var (dict, vs[i].last) + : first); + if (!last) + { + msg (SE, _("%s is not a variable name."), vs[i].last); + goto error; + } + + size_t first_idx = var_get_dict_index (first); + size_t last_idx = var_get_dict_index (last); + if (last_idx < first_idx) + { + msg (SE, _("%s TO %s is not valid syntax since %s " + "precedes %s in the dictionary."), + vs[i].first, vs[i].last, + vs[i].first, vs[i].last); + goto error; + } + + for (size_t j = first_idx; j <= last_idx; j++) + { + struct variable *v = dict_get_var (dict, j); + if (opts & PV_NUMERIC && !var_is_numeric (v)) + { + msg (SW, _("%s is not a numeric variable."), var_get_name (v)); + goto error; + } + + if (*n_vars >= allocated_vars) + *vars = x2nrealloc (*vars, &allocated_vars, sizeof **vars); + (*vars)[(*n_vars)++] = v; + } + } + + return true; + +error: + free (*vars); + *vars = NULL; + *n_vars = 0; + return false; +} + /* A set of variables. */ struct var_set { + bool names_must_be_ids; size_t (*get_cnt) (const struct var_set *); struct variable *(*get_var) (const struct var_set *, size_t idx); bool (*lookup_var_idx) (const struct var_set *, const char *, size_t *); @@ -729,6 +887,12 @@ var_set_destroy (struct var_set *vs) if (vs != NULL) vs->destroy (vs); } + +static bool +var_set_get_names_must_be_ids (const struct var_set *vs) +{ + return vs->names_must_be_ids; +} /* Returns the number of variables in VS. */ static size_t @@ -778,6 +942,7 @@ struct var_set * var_set_create_from_dict (const struct dictionary *d) { struct var_set *vs = xmalloc (sizeof *vs); + vs->names_must_be_ids = dict_get_names_must_be_ids (d); vs->get_cnt = dict_var_set_get_cnt; vs->get_var = dict_var_set_get_var; vs->lookup_var_idx = dict_var_set_lookup_var_idx; @@ -854,6 +1019,7 @@ var_set_create_from_array (struct variable *const *var, size_t var_cnt) size_t i; vs = xmalloc (sizeof *vs); + vs->names_must_be_ids = true; vs->get_cnt = array_var_set_get_cnt; vs->get_var = array_var_set_get_var; vs->lookup_var_idx = array_var_set_lookup_var_idx; @@ -891,7 +1057,7 @@ lex_match_variable (struct lexer *lexer, const struct dictionary *dict, const st *var = parse_variable_const (lexer, dict); - if ( *var == NULL) + if (*var == NULL) return false; return true; } @@ -924,15 +1090,15 @@ parse_internal_interaction (struct lexer *lexer, const struct dictionary *dict, *iact = NULL; return false; } - + assert (v); - if ( *iact == NULL) + if (*iact == NULL) *iact = interaction_create (v); else interaction_add_variable (*iact, v); - if ( lex_match (lexer, T_ASTERISK) || lex_match (lexer, T_BY)) + if (lex_match (lexer, T_ASTERISK) || lex_match (lexer, T_BY)) { return parse_internal_interaction (lexer, dict, iact, iact); }